Abstract: This article uses a two-period version of the standard economic model of deterrence to study whether sanctions should depend on an individual's record of prior convictions -- his offense history. The principal contribution of the article is to demonstrate that it may be optimal to treat repeat offenders disadvantageously because such a policy serves to enhance deterrence: When an individual contemplates committing an offense in the first period, he will realize that if he is caught, not only will he bear an immediate sanction, but also -- because he will have a record -- any sanction that he bears in the second period will be higher than it would be otherwise.